Who will be America's next top model? Ask Instagram

Beauty may be in the eye of the beholder, but Instagram could hold the key to predicting who will be the next top model in the fashion world.

Using both physical and professional information about fashion models, along with Instagram data collected in the fall of 2014, scientists from Indiana University created a formula that they say successfully forecast most of the popular new models who appeared on Fashion Week runways.

Popularity was defined as the number of runway walks that a new model participated in during the Fall/Winter 2015 season.

Marc Jacobs F/W 15 💣 @dnamodels @establishmentny @kegrand are unreal ❤️

A photo posted by melanie culley (@babesintoyland) on

To conduct the study, the researchers gathered statistics on 400 models from the Fashion Model Directory -- a major database of professional female models. They tracked the models' hair and eye color, height, hip, waist, dress and shoe size.

Next, the team looked at which agencies the models worked for.

"Models with a top agency have, everything else being equal, nearly ten time higher chances of walking a runway than their counterparts represented by non top agencies," the study authors wrote.

Third, the scientists turned their attention to the models' Instagram accounts. They noted the number of posts, "likes," and comments for each account. They also noted whether these comments were generally positive or negative.

"When we added the social information, we realized that we will be able to predict with above 80 percent accuracy whether a new face, a new model that just started ... would become popular, would run some top runway in the immediate future," Emilio Ferrara, a computer scientist at the University of Southern California, told CBS News. Ferrara, who was formerly at the School of Informatics and Computing at Indiana University, conducted the research.

💋

A photo posted by Arina Levchenko (@levchenko_arina_) on

To test their ability to predict the most popular models of the Fall/Winter 2015 season, the researchers narrowed their focus to 15 "new faces" -- models who were just starting out. Of the eight models the researchers predicted would become most popular, six did.

They were Sofia Tesmenitskaya, Arina Levchenko, Renata Scheffer, Sasha Antonowskaia, Melanie Culley and Phillipa Hemphrey.

Conversely, of the seven models who researchers predicted to score lowest in popularity, six did.

The researchers also looked at the Instagram accounts of the more established models. Here they found that buzz on social media could help predict their success on the runway, too.

"Being more active played in your favor," said Ferrara. "If you're a model and you're on Instagram, you are already more likely to become famous than one of your colleagues who was not on Instagram. So already being there plays in your advantage."

The study found that a high number of "likes" and comments, as well as frequent posting, pointed to success on the runway. The tone of the comments did not affect popularity. Models with a higher than average number of Instagram posts had a 15 percent higher chance of walking a runway. But surprisingly, more "likes" could lower those chances by about 10 percent.

"Our analysis suggests that Instagram is as important as being cast by a top agency in terms of its ability to predict success on the runway," Ferrara said in a press release. In addition, he told CBS News, "physical attributes alone don't determine success."

Still, physical attributes certainly help. The study found that being an inch taller than the average for the group roughly doubled a model's chances of walking a runway. Those with larger dress, hip and shoe sizes than other models had less chance of runway success.

f

We and our partners use cookies to understand how you use our site, improve your experience and serve you personalized content and advertising. Read about how we use cookies in our cookie policy and how you can control them by clicking Manage Settings. By continuing to use this site, you accept these cookies.